This invention concerns light source equipment that includes what is called a dielectric-barrier discharge lamp, which is a type of discharge lamp used, for example, as a source of ultraviolet radiation for photochemical reactions, in which eximer molecules are formed by dielectric-barrier discharge, and which uses light emitted from the excimer molecules.
Technical literature explaining the technology involved in the dielectric-barrier discharge lamps with which this invention is concerned can be found in, for example, JPO kokai patent report H2-7353. This document describes an emitter that produces light by causing the formation of eximer molecules by means of a dielectric barrier discharge in a discharge chamber filled with a discharge gas that forms eximer molecules, and using the light radiated by those eximer molecules (a dielectric barrier discharge is also known as ozonizer discharge or silent discharge; see Denki Gakkai, xe2x80x9cDischarge Handbook,xe2x80x9d revised edition, 7th printing, June 1989, p. 263).
Dielectric-barrier discharge lamps have a discharge plasma space and one or two dielectrics sandwiched between electrodes. FIG. 19(a) shows a dielectric-barrier discharge lamp 1 with two dielectrics 5, 6. In FIG. 19(a), by the way, the lamp seal 7 also serves as two dielectrics 5, 6.
When lighting up the dielectric-barrier discharge lamp 1, a high-frequency, alternating current of, for example 10 to 200 kHz and 2 to 10 kV is imposed on the electrodes 3,4.
However, because of the dielectrics 5,6 between the discharge plasma space 2 and the electrodes 3,4, current does not flow directly from the electrodes 3,4 to the discharge plasma space 2; the current flows by means of the action of the dielectrics 5,6 as a condenser. In other words, a charge equal in size and opposite in sign to that on electrodes 3,4 is induced on the discharge plasma space side of the dielectrics 5,6 because of polarization of the dielectric. The discharge occurs between the dielectrics 5,6 that face across the discharge plasma space 2.
Little current flows along the discharge plasma space 2 side of the dielectrics 5,6; when discharge occurs, the charge induced on the discharge plasma space 2 side of the dielectrics 5,6 is neutralized by the charge moved by the discharge, and the electrical field within the discharge plasma space 2 is reduced. For that reason, the current stops even if the voltage continues to be impressed on the electrodes 3,4. But when the voltage impressed on the electrodes 3,4 rises again, the discharge current continues.
When the discharge ceases after having occurred, there is no further discharge until the polarity of the voltage impressed on the electrodes 3,4 has reversed.
In the case of a dielectric-barrier discharge lamp in which xenon gas, for example, is sealed, the xenon gas is dissociated into ions and electrons by the discharge, and becomes xenon plasma. When the xenon plasma is excited to a specified energy level, eximer molecules are formed within the plasma. Xenon eximers divide after a certain lifespan, but the energy released at that time is emitted as a photon of vacuum ultraviolet wavelength. To make a dielectric-barrier discharge lamp work efficiently as a vacuum ultraviolet light source, it is necessary to form the eximer molecules efficiently.
The greatest obstacle to efficient formation of eximer molecules during discharge is the excitation of the discharge plasma to energy levels that do not contribute to the formation of eximer molecules.
The movement of discharge plasma electrons immediately after discharge begins is collective, and the energy is high but the temperature is low. In this state, the discharge plasma has a high probability of transition to the resonant state required for formation of eximer molecules. If the discharge time is prolonged, however, the movement of the plasma electrons gradually becomes thermal. That is, it reaches a state of thermal equilibrium known as a Maxwell-Boltzmann distribution; the plasma temperature rises, and there is an increased probability of transition to a state of higher excitation where eximer molecules cannot form.
Moreover, sometimes when eximer molecules have been formed, a subsequent discharge will break down the eximer molecules before their lifespan elapses and they divide naturally by emitting the desired photon. In fact, in the case of xenon eximers, a period of about 1 xcexcs is required between the beginning of discharge and emission of a vacuum ultraviolet photon, and a subsequent discharge or redischarge during that period reduces the efficiency of eximer light emission.
In other words, once discharge had commenced, it is most important to reduce as much as possible the energy of subsequent discharges.
Even in the event that the discharge time is short, if the energy injected during the discharge period is too great, there is similarly an increased probability of transition to a state of high excitation. Plasma that has transitioned to a state of high excitation moderates itself by emission of infrared radiation, which just raises the temperature of the lamp and does not contribute to eximer light emission.
That is, the discharge must be driven so as to suppress the excitation of discharge plasma to energy levels that do not contribute to the formation of eximer molecules. That point is one that cannot be satisfied by conventional dielectric-barrier discharge lamp light source equipment.
JPO kokai patent report H1-243363 is a proposal to achieve eximer light emission with high efficiency by means of all pulse discharges, including dielectric-barrier discharges. This follows the condition stated above that once a discharge has begun, the energy of the subsequent discharge is reduced as much as possible. However, what is described in this proposal is which parameters to control to increase the efficiency of eximer light emission; there is no specific mention of the effective conditions for those parameter values. Particularly in the case of dielectric-barrier discharges, there is little freedom for control of the voltage that has to be impressed and the current that has to be injected into the discharge plasma space through the dielectric, and it is extremely difficult to discover the optimum conditions.
There are proposals to improve the efficiency of dielectric-barrier discharge lamps, such as JPO patent report H8-508363 (U.S. Pat. No. 5,604,410). However, these reports say nothing about specific items that are truly effective in achieving control of the excitation of discharge plasma to energy levels that do not contribute to formation of eximer molecules, so as to form eximer molecules efficiently. There are descriptions of the results of experiments on variation of brightness relative to variation of impressed voltage, in connection with frequency and duty cycle, with regard to short pulse arrays and short waveforms, as well as explanations of improvement of efficiency relative to conventional sinewave drive.
Actual power supplies, however, include high-voltage transformers, and are unable to impress ideal short pulse arrays and short waveforms; because of interaction of the output impedance of the power supply and the impedance of the lamp, the waveform lacks precision, and partial resonance causes a sinewave voltage to be impressed.
In the event of discrepancies from ideal short waveforms in such actual power supplies, unless it is clear what degree of discrepancy is allowed, it is not possible to design or manufacture practical light source equipment economically.
This invention was made in consideration of the situation described above; its purpose is provide dielectric-barrier discharge lamp light source equipment that can produce eximer molecules efficiently and operate efficiently as a vacuum ultraviolet light source.
In order to form eximer molecules efficiently, which is the task of the invention of this application, the excitation of discharge plasma at energy levels that will not contribute to the formation of eximer molecules is controlled. To accomplish that, once the voltage impressed on the lamp has climbed at a limited rate of ascent, reached the voltage for the commencement of discharge and started the discharge, then the discharge should be stopped as quickly as possible.
Operation of the electrical circuitry of the dielectric-barrier discharge lamp 1 is explained by FIG. 19(b). As shown in FIG. 19(b), the discharge path of the discharge plasma space 2 continues in series with a resistor 10 and a switch 11. The dielectric-barrier discharge lamp 1 has dielectrics 5,6 between the electrodes 3,4 and the discharge plasma space 2, and it functions as a condenser within the electrical circuitry. In the event that there arc two pieces of dielectric, however, the two condensers can be thought of as a single condenser 13.
Because the structure has this condenser inserted in series with the discharge plasma space 2, discharge current flows through the dielectric-barrier discharge lamp 1 only for a period immediately after the change in polarity of the voltage impressed on the lamp, and a non-discharge period occurs naturally, even without impressing voltage on the lamp as a pulse voltage that has a rest period that is essentially zero voltage.
Moreover, discharge does not occur until the voltage of the discharge plasma space 2 reaches the voltage for the commencement of discharge.
The discharge plasma space 2 itself forms a condenser 12, and when discharge begins, almost all the energy stored in this condenser is expended, and so once the discharge begins, there is no need for additional current to the dielectric-barrier discharge lamp 1 from the power supply.
The unit area of the lamp wall surface is considered next.
The voltage for the commencement of discharge is decided almost automatically when the gas pressure and the discharge gap are decided. And because the size of the discharge gap determines the static capacity C1 of the condenser 12 formed by the discharge plasma space, the minimum energy that can be put into the plasma between the commencement and completion of a single discharge is the energy of a full discharge of the charge stored in the condenser 12 formed by the discharge plasma space. That determines the constitution of the lamp.
It was stated above that in order to form eximer molecules efficiently, the excitation of discharge plasma at energy levels that will not contribute to the formation of eximer molecules is controlled. That control is best achieved through the conditions for discharge of that minimum energy.
However, the conditions for discharge of that minimum energy can be realized in an ideal fashion by very slowly raising the voltage impressed on the lamp, and then discharging it, using a power supply with an output impedance that is quite large.
There are, however, problems in applying such a power supply in an actual light source.
The first problem is that when the output impedance is large, it is not possible to operate fast enough to discharge in cyclical repetitions.
The second problem is that under conditions for discharge of the minimum energy, there will be a lack of uniformity of discharge within a single lamp because of a lack of positional uniformity of the discharge gap within the lamp.
Accordingly, using a power supply that has the small output impedance that can realize the necessary volume of light, in order to have a practical light source that produces uniform discharges over the full surface of the dielectric-barrier discharge lamp, it is necessary to raise the voltage impressed on the lamp beyond the minimum energy conditions for discharge. However, the degree to which the voltage impressed on the lamp is raised should be held below the range that would allow a drop in the efficiency of eximer light emission.
That is, the peak value of the voltage impressed on the lamp should be based on the lowest value at which the lack of discharge uniformity can be ignored practically, and should be held to 2 times that lowest value, preferably 1.5 times the lowest value. Or to base it on the lowest value for maintaining the discharge, it should be held to 3 times, preferably 2.5 times that lowest value. Now, when the lamp power is increased, it is done not by raising the voltage impressed on the lamp, but by increasing the drive frequency of the lamp power supply. Because a fixed volume of lamp current flows each time the voltage impressed on the lamp is reversed, the lamp power is proportional to the drive frequency. Therefore, by increasing that drive frequency, it is possible to increase the lamp power without the bad effects that would accompany an increase of the voltage impressed on the lamp.
The basic constitution of the dielectric-barrier discharge lamp light source equipment is shown in FIG. 1. In the figure, 1 is the dielectric-barrier discharge lamp described above, and 8 is the power supply. The power supply 8 comprises a high-frequency alternating current power supply 9 of the bridge, push-pull or fly-back type, and a step-up transformer Tr.
Typical waveforms for the voltage impressed on the lamp E(t), the discharge gap voltage (that is, the voltage of the discharge plasma space) Vl(t), lamp current Is(t) and discharge current Id(t) of a practical power supply are shown in FIG. 2 (the figure shows waveforms obtained from a computer simulation modelling the circuit and lamp based on a full-bridge power supply, but they would be basically the same with a half-bridge or push-pull type power supply).
Now with regard to the discharge current Id(t), it is the current in the lamp flowing through the resistor in FIG. 19(b), so the waveform cannot be measured directly, but if it is possible to measure the waveform data for the voltage impressed on the lamp E(t) and the lamp current Is(t), it can be calculated from the static capacitance C1 of the condenser 12, which is the discharge plasma space shown in FIG. 19(b), the static capacitance C2 of the condenser 13 which is the dielectrics 5,6, and floating static capacitance C3 which exists in parallel with the dielectric-barrier discharge lamp.
That is, the static capacitance C1 of the condenser 12, which is the discharge plasma space, the static capacitance C2 of the condenser 13 which is the dielectrics 5,6, and floating static capacitance C3 which exists in parallel with the dielectric-barrier discharge lamp determine the two relationships F=1+C1/C2 and Cv=C1+C3xc2x7F. Using that, the discharge current Id(t) can be found from the following formula (1).
Id(t)=Fxc2x7Is(t)xe2x88x92Cvxc2x7dE(t)/dtxe2x80x83xe2x80x83(1)
Because this method uses numerical differentiation, the precision is not to good in regions where the current values of the waveform obtained as a result are small, but it does indicate the rapid rise when discharge commences, so there is no problem as long as it is used for that purpose.
In FIG. 2, when the voltage impressed on the lamp E(t) suddenly changes polarity, the discharge gap voltage Vl(t) also makes a quick change, and when it reaches the voltage for the commencement of discharge at point G1 discharge commences. When discharge commences, the discharge current waveform J1 (see the discharge current Id(t) waveform in FIG. 2), rapidly appears, and as a result the discharge gap voltage Vl(t) drops suddenly. In response to this sudden drop of the discharge gap voltage Vl(t) (that is, the voltage in the discharge plasma space), the voltage impressed on the lamp E(t) also drops, producing the indentation at point K.
The point on the voltage impressed on the lamp waveform that corresponds to point J2 where the discharge current ceases is the point P1 with the largest absolute value, or slightly beyond that. In evaluating actual light source equipment, that point can be taken as located at the point P1 with the largest absolute value.
Thereafter, the change of the waveform of the voltage impressed on the lamp is reduced by C2/(C1+C2) times, and the waveform of the discharge gap voltage Vl(t) is expressed as is. As used here, C1 and C2 are the static capacitance C1 of the condenser 12, which is the discharge plasma space, and the static capacitance C2 of the condenser 13 which is the dielectrics 5,6, as described above. In the case of a dielectric-barrier discharge lamp that has dielectrics 5,6 for both electrodes 3,4, C2 can be thought of as the independent static capacitance of each dielectric, in series.
In a practical power supply, the voltage impressed on the lamp has an indentation at point K for the following reason. In response to the sudden drop of the discharge gap voltage Vl(t) (which is the voltage of the discharge plasma space) described above, the voltage impressed on the lamp E(t) drops as well. The power supply 8 attempts to compensate for the drop in the voltage impressed on the lamp, but because of the presence of an inductive output impedance caused by the magnetic leakage of the step-up transformer or the inductance of the cable, compensation for the drop in the voltage impressed on the lamp is delayed, and as a result, the indentation in the direction of the greatest absolute value at point K is produced. Moreover, because of resonance between this inductive output impedance and the static capacitance of the dielectric-barrier discharge lamp 1, an oscillating component is incorporated in the voltage impressed on the lamp following the indentation at point K. The points with greatest absolute value and least absolute value are produced on the waveform of the voltage impressed on the lamp as a result.
Hereafter, the oscillation that follows the indentation at point K is called ringing, and its frequency is the ringing frequency Fr. The time period from the commencement of discharge at point K to the peak value of the voltage impressed on the lamp E(t) is called xcfx84, and the period from the peak P1 that appears after the indentation at point K to the next peak P2 is T12.
This invention focuses on variation of the waveform of the voltage impressed on the lamp following the indentation at point K, and attempts to inject power into the dielectric-barrier discharge lamp effectively, without lowering the efficiency of ultraviolet light emission.
That is, eximer molecules are produced by the discharge that occurs with the sudden rise or fall of the waveform of the voltage impressed on the lamp E(t) as described above, and the eximer molecules are broken down if further discharge current is added in the period before the eximer molecules dissociate and produce ultraviolet radiation. Accordingly, it is desirable from the perspective of efficient production of ultraviolet radiation to cut off the discharge current quickly.
Nevertheless, the discharge current continues from the commencement of discharge at the indentation at point K to the first peak of the voltage impressed on the lamp E(t). Therefore, if there is a long period xcfx84 from the commencement of discharge until the voltage impressed on the lamp peaks, the efficiency of ultraviolet radiation is reduced as a result. Moreover, in the state where ringing occurs following the indentation at point K, the discharge current continues until the peak is reached. Thus when the ringing frequency is low, it takes longer for the discharge current to cease, and the efficiency of ultraviolet radiation is reduced as a result.
In other words, a dielectric-barrier discharge lamp can be made to discharge without reducing the efficiency of ultraviolet light emission by shortening the period xcfx84 from the commencement of discharge to the first peak of the voltage impressed on the lamp, or by increasing the ringing frequency Fr.
In this case, the period xcfx84 from the commencement of discharge to the first peak of the voltage impressed on the lamp and the ringing frequency Fr are generally determined by the impedance L and the static capacitance C of the circuit comprising the power supply 8 and the dielectric-barrier discharge lamp 1. By making these values smaller, it is possible to shorten the time xcfx84, and to increase the ringing frequency.
In this case the discharge ceases when the first peak P1 is passed (that is, the switch in FIG. 19(b) is opened at that time), so when the first peak P1 is passed, the waveform of the voltage impressed on the lamp E(t) oscillates up and down at a frequency determined by the static and floating capacitances of the condensers 12,13 and the circuit inductance L show in FIG. 19(b).
Because the discharge is completed immediately after the first peak P1 is passed, the period T12 from the first peak P1 to the second peak P2 will correspond to the period of oscillatory frequency determined by the static and floating capacitances of the condensers 12,13 and the circuit inductance L.
Generally it is possible to calculate the resonant frequency of an LC resonant circuit by the following formula:
LC=1/(2 xcfx80f)2
Accordingly, given the inductance L and static capacitance C of the circuit above, the ringing frequency Fr is roughly determined by 1/{2 xcfx80xc3x97f(LC)} and the value of the static capacitance C depends on the static capacity of the dielectric-barrier discharge lamp 1, so to increase the ringing frequency Fr (shorten the period xcfx84) it is best to reduce the value of the inductance L. Specifically, it is possible to increase the ringing frequency Fr by reducing the coupling inductance of the step-up transformer Tr.
On the basis of the concept stated above, the inventors investigated the relationship between the period xcfx84 and the efficiency of ultraviolet light emittance during period T12 from peak P1 to peak P2.
As a result, it was learned that, as shown in FIGS. 8, 9 and 10 described below, it is possible to discharge a dielectric-barrier discharge lamp without reducing the ultraviolet light emission efficiency if the period xcfx84xe2x89xa62.1 xcexcs, the period T12xe2x89xa63 xcexcs, or the ringing frequency Frxe2x89xa7300 kHz.
In this case, the ringing frequency Fr is roughly determined by 1/{2 xcfx80xc3x97(LC)}, so it is best if LC less than 2.8xc3x9710xe2x88x9213, where the unit of C is F (farad) and the unit of L is H (henry).
Because the value of LC stated above is roughly determined by the static capacitance of the lamp 1 and the inductance of the transformer Tr, the inductance of the transformer Tr should be selected so that the value of LC meets the condition stated above with respect to the static capacity of the lamp 1.
The measured value of the static capacitance C can be gotten from such things as the inductance while the lamp is burning. The measurement of L should be gotten by measuring the inductance on the secondary side of the transformer Tr with an impedance meter while recreating the state on the primary side based on the state immediately after the completion of discharge when lighting the lamp.
If a full-bridge or half-bridge high-frequency alternating current power supply is used, for example, the inductance on the secondary side of the transformer can be measured while the primary side is shorted. In the case of a push-pull power supply, the inductance on the secondary side of the transformer can be measured by shorting the center of the primary side with one end, leaving the other end open. And in the case of a fly-back power supply, the inductance on the secondary side of the transformer Tr can be measured with the primary side open.
Now, depending on the power supply 8 being used, sometimes the indentation at point K mentioned above will not appear clearly in the waveform of the voltage impressed on the lamp.
The absence of a clear indentation in the waveform of the voltage impressed on the lamp is particularly frequent in the case of a fly-back power supply. In that case, the waveform of the discharge current Id(t) can be calculated, and the steepest part of its rise taken as the indentation.
The invention described in claims 1 to 3 of this invention is one that defines the practical conditions for discharge in a dielectric-barrier discharge lamp with good efficiency, based on what is stated above; by fulfilling any of the following conditions, it is possible to make a dielectric-barrier discharge lamp discharge efficiently.
(1) After passing the point in time when the voltage value to start a dielectric-barrier discharge is reached, voltage is impressed on the dielectric-barrier discharge lamp for a period of not more than 3 xcexcs from the largest voltage value to the next largest voltage value.
(2) Voltage is impressed on the dielectric-barrier discharge lamp for a period of not more than 2.1 xcexcs from the voltage value at which the dielectric-barrier discharge commences until the maximum voltage value in the next round of voltage impression.
(3) Taking L as the inductance and C as the static capacitance of the circuit comprising the supply equipment and the dielectric-barrier discharge lamp, the inductance L and static capacitance C are selected to fulfill the following formula:
LCxe2x89xa62.8xc3x9710xe2x88x9213